Electronic structure of tris(2-phenylpyridine)iridium: electronically excited and ionized states

A computational study of tris(2-phenylpyridine)iridium, Ir(ppy)(3), is presented. The perspective is that of using organo-transition-metal complexes as phosphorescent species in light-emitting diodes (OLED's). Quantum yields approaching 100% are possible through a triplet harvesting mechanism. Complexes such as Ir(ppy)(3) are amenable to exacting experimental and theoretical studies: small enough to accommodate rigor, yet large enough to support bulk phenomena in a range of host materials. The facial and meridional isomers differ by similar to 220 meV, with fac-Ir(ppy)(3) having the lower energy. Because fac-Ir(ppy)(3) dominates in most environments, focus is on this species. Time-dependent density functional theory using long-range-corrected functionals (BNL and omega B97X) is used to calculate excited states of Ir(ppy)(3) and a few low energy states of Ir(ppy)(3)(+). The calculated T-1-S-0 energy gap (2.30 eV) is in reasonable agreement with the experimental value of 2.44 eV. Only a few percent of singlet character in T-1 is needed to explain so short a phosphorescence lifetime as 200 ns, because of the large (LC)-L-1 <- S-0 and (MLCT)-M-1 <- S-0 absorption cross-sections. Equilibrium geometries are calculated for S-0, T-1, and the lowest cation state (D-0), and several ionization energies are obtained: adiabatic (5.86 eV); vertical from the S-0 equilibrium geometry (5.88 eV); and vertical ionization of T-1 at its equilibrium geometry (5.87 eV). These agree with a calculation by Hay (5.94 eV), and with the conservative experimental upper bound of 6.4 eV. Molecular orbitals provide qualitative explanations. A calculated UV absorption spectrum, in which transitions are vertical from the S-0 equilibrium geometry, agrees with the room temperature experimental spectrum. This is consistent with Franck-Condon factors dominated by Delta v(i) = 0, as expected given the delocalized nature of the orbitals. Ir(ppy)(3) vibrational frequencies were calculated and used to estimate the probability density P(E-vib) for 500 K, i.e. the temperature at which the experiments were carried out. In combination with the vibrational energy imparted through (LC)-L-1 <- S-0 photoexcitation, it is seen that a large amount of vibrational energy appears in Ir(ppy)(3)(+) without causing its fragmentation. Specifically, for hv - E-T1 = 15,000 cm(-1), the probability density for total vibrational energy peaks at similar to 31,000 cm(-1) with a 7800 cm(-1) width.